The smartphone killer
The Humane AI Pin illustrates another trap: the obsession with total autonomy. By trying to pack all the compute into the wearable to make it a standalone device, its creators marketed it as the phone's replacement. That overlooks the fusional bond people have with their smartphones; any attempt at abrupt substitution is perceived as an attack. The paradigm shift will never come from aggressive replacement marketing, but from a gentle transition: the gradual migration of the phone's features to new peripherals.
Paradoxically, the rise of the smartphone was built on absorption: it swallowed our radios, MP3 players, alarm clocks, cameras and notebooks. Tomorrow, voice interfaces and peripheral AI will nibble away at navigation and communication, effectively offloading screen use. Other features, such as photography, will remain the phone's preserve for many decades to come.
Trying to design an architecture that is completely independent of the smartphone, especially at the dawn of a new product category, is a strategic mistake. The “smartphone killer” promise systematically has two harmful effects at launch: it captivates the market by setting artificially high expectations, then frustrates users when they hit the inevitable limitations, deepening the disillusionment.
The best approach is to rely on the phone for everything it does best (compute, connectivity, power) and to replace only what voice AI can transcend: the interface. In this hybrid architecture, the smartphone becomes a battery-powered “compute module”. It stays in our pocket, screen off, communicating wirelessly with the peripherals we wear. This model covers 70% of our daily digital interactions, while keeping the phone available, ready to be unlocked for complex visual tasks.
Delegating compute to the smartphone also removes a constraint that is vital to wearable adoption: design. Without a heavy processor or a massive battery, these objects can finally fade into the background. The history of technology shows that wearable devices only gain social acceptance when they look like objects we already wore (glasses, watches, jewellery) or when they become invisible. Reducing the wearable to its simplest input/output expression is the only path to that essential discretion.
Next article: "Architecture choices in wearables" ← Back to the blog
